Just because you speak English it doesn’t mean that your English sounds the same compared to the person standing next to you. This is because there are variances like the timbre in your voice, your upbringing, your gender, and of course things like race and heritage play a part in the way you sound.
This clearly presents as a challenge when developing voice-based software in which developers need to ensure that their software can understand as many accents as possible. That being said, which software works the best? That’s what the folks at WIRED tried to found out in a test in the video above, where they took a bunch of participants with various accents and tested them out on the Google Assistant, Amazon Alexa, and Apple’s Siri.
For anyone who’s used Google Now or Google Assistant, it probably shouldn’t come as a surprise to learn that Google Assistant came out on top in the test. Participants were told to deliberately mispronouncing words and names to see if the software could catch on, and for the most part Google Assistant held its own and stumbled a bit when it came to Italian accents.
Apple’s Siri came in second while Amazon’s Alexa platform performed the worst of the three, although we guess to be fair Amazon’s entrance into the voice recognition software was recent, so perhaps with time we can expect improvements to be made to Alexa.